Background On Dec 1, 2006, Mexico's public health-care insurance scheme, Seguro Popular, implemented the Medical Insurance Century XXI (SMSXXI) programme, to provide insurance to children younger than 5 years without social security. SMSXXI aims to increase access to health services, decrease out-of-pocket health expenses (OOPHE), and reduce health inequities. SMSXXI covers uninsured, primarily low-income, populations who might be most at risk of the financial and health consequences of costly medical interventions. Methods We assessed the effects of SMSXXI on health outcomes and financial protection for Mexico's children using multiple nationally representative surveys and administrative data sources spanning 2001-16. The identification of effects relied on detailed hospital-level affiliation data mapping the geographical expansion of SMSXXI's coverage across the country over time. The units of analysis included hospitals, households, and children. Primary outcomes were neonatal and infant mortality, self-reported morbidity (health status, influenza, and diarrhoea), and child's height. Secondary outcomes were OOPHE, hospital discharges, and quality of service provision. Effects controlled for fixed and time-variant confounders using double-difference and triple-difference estimation strategies. Where feasible, we also estimated effects using exogenous variation in programme eligibility rules that limited enrolment in SMSXXI to children born after Dec 1,2006. Findings SMSXXI was not associated with early (<1 week) neonatal mortality, but was associated with a reduction in late (<28 days) neonatal mortality by 0.139 deaths per 1000 livebirths (95% CI 0.032-0.246), or 7% (2-12) relative to the comparison base of 1.98 deaths per 1000 livebirths in 2006. SMSXI was associated with a reduction in infant mortality from conditions covered by the programme by 0.147 deaths per 1000 livebirths (0.023-0.271), or 5% (1-10) relative to the comparison base of 2.73 deaths per 1000 livebirths. The effects were largest in high baseline mortality areas. Long-term health effects, 8 years after the onset of SMSXXI, were reflected in a 0.434 cm (0.404-0.459) height increase for birth cohorts exposed to the programme and an average effect on height of 0.879 cm (0.821-0.932) for low-income populations. About 3-6 years after SMSXXI started, children reported having better health status and lower incidence of influenza and diarrhoea. The programme led to a 14% reduction (7-28) in OOPHE, primarily from hospital-related expenses. No effects were detected on hospital discharges, suggesting that SMSXXI might not have increased use. Interpretation SMSXXI promoted access to covered interventions and encouraged better primary care. The programme also promoted increased supply and quality of care by improving human and physical resources sensitive to unmet needs. Increased resource availability and improved supply of health care, rather than increased use, contributed to reduce infant mortality and improved long-term health as proxied by self-reported morbidity and child height. Consistent with the programme's focus on uninsured and low-income populations, the effects on mortality, long-term health status, and OOPHE were concentrated in vulnerable groups. Copyright (C) 2019 The Author(s). Published by Elsevier Ltd.
